“I prefer go”:
English L2 Verb Complement Errors
Mary Lou Vercellotti, University of Pittsburgh
Nel de Jong, Queens College of CUNY
Infinitival and gerundive verb complements can pose difficulties for L2 learners
because English verbs vary on which type of complement they require or allow.
Some take only a gerund (I enjoy walking) or an infinitive (I want to walk) while
others allow both forms (I like walking/to walk.). With no straightforward “rule” to
apply, English L2 learners must memorize each individual verb with the
acceptable verb complement(s).
This poster describes the use and accuracy of verb complements in over 200
minutes of recorded speech from a training task in which students spoke about a
given topic three times (Nation, 1989). Participants were 23 high-intermediate
students in an intensive English program in a large university. They had explicit
knowledge of verb complements from their regular language classes, but the
instruction in this study was implicit, since no instruction was given to produce or
to attend to the target structure.
Verbs that could take both structures were the most common (67.4%) but the
least accurate. It appears the students had more difficulty choosing which verb
complement form to use when both were allowed. This suggests that the
variability of input impedes acquisition. As a result, students often produced only
the base form of the verb complement (* I like walk), which accounted for 72% of
the errors in this category. It is argued that the students were not always able to
apply their explicit knowledge correctly in a spontaneous production task.
Verb Complement Forms Produced
Gerund (n =177)
Infinitive (n = 294)
* to V -ing (n = 15)
* Neither (n = 55)
• Students produced more infinitival than gerundival verb
• 13% of verb complements were ambiguously produced
Accuracy in Verb Complement Structures
English Verb Complement Structure
• Some matrix verbs only allow a gerund
I enjoy watching
• Some matrix verbs only allow an infinitive
I want to study
• Some matrix verbs allow variation
I like playing / to play
Accuracy by Matrix Verb
Previous explicit verb complement instruction in grammar class.
• Verbs were grouped by acceptable verb complements
verb + infinitives
verb + gerund OR infinitive
My friend agreed to go hiking I can’t stand waiting in line, =
with me.
I can’t stand to wait in line.
begin, continue, hate, like, love,
prefer, start
• There is a high frequency of learner-centered matrix verbs that can
take verb complements in the speaking class textbook series,
especially considering the low number of verbs (Juffs, 1998).
• Students have exercises to fill in the correct form (gerund or infinitive)
for the given matrix verb.
Data was collected during a production task in which students spoke of
a topic for 4 minutes, then again in 3 minutes, and then in 2 minutes
(Nation, 1989).
• No instruction given to produce or attend to verb complement
• This production task is assumed to require implicit knowledge.
• Students gave 3 speeches per session x 3 sessions in one month
Nominative non-finite verb complements were analyzed.
Can these English L2 learners spontaneously produce correct
verb complement structures?
Are errors a result of the idiosyncratic requirement of the matrix
• The number of verb complements were topic dependent; a
comparison between topic/session is not relevant.
• The students’ mean accuracy average was 82.2% (s.d.12%)
• Few instances of variation of structure from recording to recording
within one activity; repetition did not seem to affect accuracy.
• Few instances of on-line corrections of the verb complement
• Two corrections (ex. like stay [*], like to stay)
• Four incorrect (ex. I like to do, do it myself , to doing [*] it myself)
Gerund w/ error
Infinitive w/ error
Both to V -ing
Participants – 23 high-intermediate English L2 (mixed L1) students
appreciate, avoid, dislike, enjoy, agree, appear, can(‘t) afford,
keep (don’t) mind, miss, practice decide, learn, mean, offer, plan,
refuse, seem, volunteer, wait
Gerund Only
verb + gerund
I enjoy watching old
2% 7%
Historical language change with gerundives increasing (Fanego, 2004)
• with infinitive entrenchment of collocations (esp. emotive verbs)
The verb complement structure is also difficult because
• Gerundival –ing has the same form as progressive –ing
• Infinitival to has the same form (and historical meaning) as the
preposition to (Duffley, 2000).
Variable Verb Complement
• Matrix verbs that allow either form were most common and had
the most errors (85% accuracy)
• When either form is allowed, the most common error was
producing only the root verb with neither morpheme (n =39)
• Two-thirds of the errors for the infinitive only matrix verbs was
producing only the root verb (n=14).
• Errors were not often a result of a mismatch between matrix verb
and verb complement.
• The most common error overall was producing only the root in
the verb complement, with neither the infinitival to marker nor the
gerundival -ing.
• These findings suggest that
• the students could not always apply their explicit knowledge of
verb complement structure during this production task, which
indicates that this structure has not been fully proceduralized
(DeKeyser, 1997)
• the variation may impede acquisition
• the increased processing demands of the verb complement
structure may result in a lack of production of the grammatical
markings (to and –ing)
• teachers may want to focus more instruction on verb
complement structure as the forms are shared with the
preposition to and the progressive –ing.
• Follow-up analysis will include the number and length of pauses
around the verb complement structure as a measure of
monitoring and possible reliance of explicit knowledge.
DeKeyser, R. M. (1997). Beyond explicit rule learning: Automatizing second
language morphosyntax. Studies in Second Language Acquisition, 19, 195221.
Duffley, P. (2000). Gerund versus infinitive as complement of transitive verbs in
English: The problems of “tense” and “control. Journal of English Linguistics
28, 221-248.
Fanego, T. (2004). Is cognitive grammar a usage-based model? Toward a
realistic account of English sentential complements. Miscelanea: A journal of
English and American studies 29. 23-58.
Juffs, A. (1998). The acquisition of semantics-syntax correspondences and verb
frequencies in ESL materials. Language Teaching Research 2, 2, 93-123.
Nation, P. (1989). Improving speaking fluency. System, 17, 377-384.
Funding for this research is provided by the National Science Foundation, Grant Number
SBE-0354420 to the Pittsburgh Science of Learning Center (PSLC, http://www.learnlab.org).

Fostering Oral Fluency in the Classroom with English